Grid computing refers to collections of computing resources from different locations that work towards a common goal. This grid basically acts like a distributed system having non-interactive workloads which involve many files. The grid will work on multiple tasks within any network and it may also work on specific applications. Grid is designed to find solutions to problems which are very large for super computers to handle. So, grid computing works like a processor architecture combining resources from different domains to attain a common objective. All computers on the grid will function together just as if it was a supercomputer. The grid will offer a multi-user infrastructure which accommodates sporadic demands of very large data processing. The grid is linked using parallel nodes which constitute a cluster that runs on operating systems, whether free software or Linux. This cluster may range from a tiny workstation to various networks. These are frequently used in web services like ATM banking, marketing research, back-office set-ups etc.